ROMANTIK. Transceiver AlgorIthms for Multihop NetworKs. Management and AdvaNced

Transcription

1 ROMANTIK ResOurce Management and AdvaNced Transceiver AlgorIthms for Multihop NetworKs Javier Fonollosa Universitat Politècnica de Catalunya javier.fonollosa

2 Partners UPC Universitat Politècnica de Catalunya Prof. Josep Vidal Department of Signal Theory and Communications UoB University of Bristol Prof. Andrew Nix Centre for Communications Research DUN Dune, Ingegneria dei Sistemi Otello Gasparini INFO Università di Roma La Sapienza Prof. Sergio Barbarossa INFOCOM Department ICOM Intracom Anastasia Andritsou Development Projects Department FLE Fujitsu Laboratories of Europe Dr. Seshaiah Ponnekanti Advanced Radio Access Systems

3 General technical objective Consider multi-hop networks for increased capacity and in systems beyond 3G with QoS support Estimated system requirements for systems beyond 3G 1 Peak user data rates: Average user data rate: Aggregate cell capacity (r=0.5 Km): Traffic demand per area: 10 to 20 Mbps 0.1 to 1 Mbps 25 to 50 Mbps (payload) ~1 to 200 Mbps/ km² For SB3G, area fitting the peak aggregate data rates is very reduced (specially for NLOS situations) which is uneconomical. LOS propagation conditions can be achieved in multihop networks. 1 Werner Mohr, Data Rates Estimates and Range Calculations for new Elements of Systems Beyond IMT-2000, Presented at WWRF WG4 meeting, June 25 th and 26 th 2002, London, at

4 General technical objective Multi-hop networks provide an economically viable solution to SB3G system requirements 1 Cell capacity per area [Mbps / km 2 ] Hata model, with wall penetration Hata model, without wall penetration Walfish-Ikegami model, with wall penetration Walfish-Ikegami model, without wall penetration ~1 to 200 Mbps/ km² traffic demand per area BPSK T = 4 Mbps BPSK T = 8 Mbps BPSK T = 25 Mbps BPSK T = 50 Mbps QPSK T = 100 Mbps 16QAM T = 200 Mbps 64QAM T = 300 Mbps (T throughput) Estimated practically achievable range for cell capacity per area Capacity per area corresponding to different aggregate cell capacity

5 Systems definition: Multihop TDD/UTRA (I) HBR LBR No HBR-UE Sync info

6 Systems definition: Multihop TDD/UTRA (I) HBR LBR No Feature 1: Coverage extension of HSDPA. 1 HBR-UE Sync info Infrastructure support: - Physical layer synchronism & slow channel allocation of relay slots - DL routing selected by RNC Scenario: Urban black spots, rural. Macrocells.

7 Systems definition: Multihop TDD/UTRA (I) HBR LBR No 1 2 LBR-UE HBR-UE Sync info 2 Feature Feature 2: 1: Coverage Coverage extension extension of for users outside HSDPA. the area covered by the infrastructure. Infrastructure support: Infrastructure - Physical layer support: synchronism & slow - Infrastructure channel allocation support of relay for last slots relay. - - Synchronism DL routing selected beyond by RNC area is provided by last relay. Scenario: Scenario: Urban Urban black black spots, spots, rural. rural. Macrocells. Macrocells.

8 Systems definition: Multihop TDD/UTRA (I) HBR HBR-UE 3 LBR No 1 2 LBR-UE HBR-UE Sync info 2 Feature 1: Coverage extension of Feature 2: 3: Coverage Interference extension reduction for users HSDPA. when outside every the user area may covered be by linked the infrastructure. through a relay connection even Infrastructure support: Infrastructure under - Physical layer support: of the required BR. synchronism & slow - Infrastructure channel allocation support: of relay for last slots relay. - Synchronism Physical DL routing layer selected beyond synchronism by RNC& slow area is channel provided allocation by last relay. of relay slots Scenario: - DL routing Scenario: Urban selected Urban black by black spots, RNC spots, rural. rural. Scenario: Urban Macrocells. Macrocells. black spots, rural. Macrocells.

9 System definition: Fixed relay for vertical handover HBR LBR LBR-UE No Feature: Single hop for vertical handover UTRA-to-Hiperlan2. Access point WLAN picocell Infrastructure support: - Physical layer synchronism (in TDD) & slow channel allocation of relay slots. Sync info Scenario: HBR of picocells, black spots.

10 Physical channel requirements per packet connection 4 Required active + relaying users per unit distance for a maximum of 5 hops for 20 users. Uniform density of users in a circular cell. Required active+relaying users per unit distance Base station Cell edge

11 Physical channel requirements per packet connection 4 Required active + relaying users per unit distance for a maximum of 5 hops for 20 users. Uniform density of users in a circular cell. Required active+relaying users per unit distance Base station Cell edge An average of 3.8 physical channels per user are required. The number of channels is larger the closer to the base station. How can we get increased capacity?

12 Rationale for capacity increase Affected links Reduction of the cell-breathing effect. UL - DL RRL: Relay-to-relay link DL: Downlink UL: Uplink

13 Reduction of the cell-breathing effect Downlink capacity in CDMA depends on the area, due to limited transmission power at the base station (in DL). Maximum path loss (db) DL in UMTS DL in the multihop UMTS Payload (kbps) Downlink vs. capacity

14 Rationale for capacity increase Affected links Reduction of the cell-breathing effect. UL - DL Improved spectral efficienty due to reduced near-far, MUD and adaptive modulation. UL - DL -RRL RRL: Relay-to-relay link DL: Downlink UL: Uplink

15 Improved spectral efficiency in DL (I) 165 Improved near-far effect among users in DL. In addition, multiuser detection in DL achieves increased orthogonality and capacity increase. Maximum path loss (db) Improved spectral efficiency in DL Payload (kbps) Downlink vs. capacity

16 Improved spectral efficiency in DL (II) Adaptive antenna technologies and advanced receivers lowers Eb/No requirements to further increase capacity in the DL and the relay-relay links. Space Time Orthogonal Block Coding with 2 TX antenna elements and 1 RX antenna element 24 active users and 32 chip length Walsh Hadamard codes.

17 Improved spectral efficiency in DL (and( RRL) (III) Adaptive modulation Find optimal waveforms to transmit information signals in time-frequency varying channels. Channel prediction is needed to determine in advance the optimum transmitted chirp signals. Multipath channel with 12 paths. The eigenvalue 1.25 of this channel has multiplicity 4 and the instantaneous frequencies are given by the red lines.

18 Rationale for capacity increase Affected links Reduction of the cell-breathing effect. UL - DL Improved spectral efficienty due to reduced near-far, MUD and adaptive modulation. Multihop reduced intercell interference. UL - DL - RRL UL - DL - RRL RRL: Relay-to-relay link DL: Downlink UL: Uplink

19 System definition: Fixed relays in TDD/UMTS Relay unit Base station User equipment

20 Power savings in structured relayed system (II)

21 Rationale for capacity increase Affected links Improved spectral efficiency due to reduced near-far, MUD and adaptive modulation. Reduction of the cell-breathing effect. UL - DL - RRL UL - DL Multihop reduced intercell interference. UL - DL - RRL Spatial reuse of relay channels within cell. RRL RRL: Relay-to-relay link DL: Downlink UL: Uplink

22 Spatial reuse of relay channels within cell Develop strategies for cross-layer optimisation: - Adaptive modulation (based on channel prediction) - Media Access Control (MAC) - Data Link Control (DLC) techniques - Radio Resource Management (RRM) tecniques - Routing algorithms with QoS-preservation

23 ROMANTIK Project duration: March 2002 to August Project Coordinator: Josep Vidal